Modeling the Global Sowing and Harvesting Windows of Major Crops Around the Year 2000

The lack of spatially detailed crop calendars is a significant source of uncertainty in modeling, monitoring, and forecasting crop production. In this paper, we present a rule-based model to estimate the sowing and harvesting windows of major crops over the global land area. The model considers field workability due to snow cover and heavy rainfall in addition to crop biological requirements for heat, chilling, and moisture. Using daily weather data for the period 1996-2005 as model input, we derive calendars for maize, rice, winter and spring wheat, and soybeans around the year 2000 with a spatial resolution of 0.5 degrees in latitude and longitude. Separate calendars for rainfed and irrigated conditions and three representative varieties (short-, medium- and long-season varieties) are estimated. The daily probabilities of sowing and harvesting derived using the model well capture the major characteristics of reported calendars. Our modeling reveals that field workability is an important determinant of sowing and harvesting dates and that multicropping patterns influence the calendars of individual crops. The case studies show that the model is capable of capturing multicropping patterns such as triple rice cropping in Bangladesh, double rice cropping in the Philippines, winter wheat-maize rotations in France, and maize-winter wheat-soybean rotations in Brazil. The model outputs are particularly valuable for agricultural and hydrological applications in regions where existing crop calendars are sparse or unreliable. Plain Language Summary This manuscript describes a numerical model to estimate location-specific sowing and harvesting dates of crops over the globe. Ten-year-long daily weather data and a few coefficients that represent the physiological characteristics of a crop (for instance, the amount of water needs to complete the life cycle of an annual crop) are only inputs to the model. Comparisons with the reported crop calendars indicate that the model well reproduces calendars of maize, rice, winter and spring wheat, and soybean around the year 2000 in major crop-producing countries. We also find that snow cover and heavy rainfall, which influence field workability but have not considered in earlier modeling, are important to estimate sowing and harvesting dates and multicropping patterns (for instance, a combination of winter and summer crops) affect the calendar of individual crops. Our findings are useful when simulating the responses of crop calendars to climate change.